Tải bản đầy đủ - 0 (trang)
1 Ruthenium(II) and Osmium(II) Complexes

1 Ruthenium(II) and Osmium(II) Complexes

Tải bản đầy đủ - 0trang

Metal Complexes of Pincer Ligands: Excited States, Photochemistry, and. . .



109



3MC



3MC



DE



DE



3MLCT



3MLCT



hn



knr



hn



knr



Fig. 19 Schematic illustration of the influence of cyclometallation on the energies of MLCT and

MC states



subsequent sections. The metal-centred oxidation process in the cyclometallated

complex 30 is shifted cathodically by 0.76 V compared to 32 (E1/2ox ¼ 0.49 and

1.25 V, respectively), an effect that can be readily understood in terms of the lower

charge on the former and the strong s-donor character of the metallated carbon.

The terpyridyl-based reduction process is, in contrast, destabilised by almost 0.4 V

(À1.61 and À1.24 V for 30 and 32, respectively). A similar trend emerges for the

osmium(II) complexes, albeit more readily oxidised than their Ru(II) counterparts.

A substantial red-shift of the lowest-energy absorption bands accompanies the

metallation: lmax ¼ 490 and 550 nm in acetonitrile solution for 32 and 30, respectively. Corresponding values for [Os(ttpy)2]2+ and [Os(ttpy)(dpyb)]+ are 667 and

765 nm, respectively. Given that the lowest-energy transition in ruthenium(II) and

osmium(II) polypyridyl complexes is normally formulated as an 1MLCT state, such

a shift is to be expected in the light of the electrochemical data, showing that the

highest-occupied metal-centred orbitals are destabilised more than the unoccupied

orbitals on the terpyridine.

The same trend is also observed in emission. The 3MLCT luminescence emission

maximum shifts from 640 to 784 nm on going from 32 to 30, with values for the

corresponding Os(II) complexes of 734 and 824 nm, respectively. An important feature

to note for the pair of ruthenium(II) complexes is that the luminescence lifetime in

solution at room temperature is increased upon metallation, contrary to what might be

anticipated on the basis of the energy gap law, whereby non-radiative decay processes

for a given type of excited state tend to increase with decreasing excited-state energy.

The values are 0.95 and 4.5 ns in CH3CN for 32 and 30, respectively. This effect can be

interpreted in terms of an increased separation, DE, between the emitting 3MLCT state

and higher-lying metal-centred states. As briefly explained in Sect. 1, metal-centred

d–d excited states can provide a deactivation pathway for 3MLCT excited states if they

are thermally accessible, i.e. if DE % kT (Fig. 19). Since cyclometallation has the effect

of both lowering the MLCT state and raising the MC states, DE is increased and nonradiative decay by this pathway is reduced. It should be noted that no such effect is

observed for the Os(II) complexes, since the MLCT states are already significantly

lower-lying in this case.



110



G.R. Freeman and J.A.G. Williams



The red-shift induced by cyclometallation in such complexes has potential in the

field of dye-sensitised solar cells (DSSCs). That the photosensitisation of a wideband-gap semiconductor could be achieved using a ruthenium(II) polypyridyl

complex was first demonstrated by O’Regan and Gr€atzel in 1991 [49]. The concept,

which exploits the MLCT nature of the excited states of such complexes as the

prelude to photo-induced charge injection, offers an alternative to bulk

heterojunction solar cells as a means of converting sunlight into electrical energy

with viable efficiencies [50]. In both types of technology, however, a major

problem facing the scientific community is the question of how to harvest the red

and NIR end of the solar spectrum. The bench-mark Ru(II) complex in DSSCs has

been [Bu4N]2[Ru(dcbpyH2)2(NCS)2] {“N719”, where dcbpy ¼ 4,40 -dicarboxy2,20 -bipyridine}. A challenge that has attracted many researchers is the development of alternatives to such dyes that do not contain NCSÀ ligands (which, in

practice, are labile, compromising device performance and longevity) and that

absorb efficiently at long wavelengths.

The possibility of using anionic cyclometallating ligands in place of the

combination of neutral bipyridines and monodentate anionic ligands such as

NCSÀ was put forward by van Koten and co-workers in 2007 [51]. They

examined complexes of the form [Ru(tpy)(NNC)]+, where NNC represents a

tridentate, cyclometallated 6-phenylbipyridine derivative incorporating one or

two CO2H pendant groups for anchoring to the TiO2 semiconductor. Spectral

coverage and photocurrent action spectra were obtained that are comparable to

that with N719 under the same conditions. A natural question that then arises is

how would the corresponding complexes incorporating dipyridylbenzene

ligands, [Ru(tppy)(NCN)]+, compare? The same researchers and, independently, Berlinguette and co-workers, have been addressing this question

recently [52, 53]. It transpires that the NCN systems are inferior to the NNC

analogues. For example, Fig. 20 shows the photocurrent action spectra of

DSSCs using the set of four complexes 33–36. Whilst the performance of the

NNC systems is superior to the bis-terpyridyl complex, that of the NCN complex

is poorer. The difference can be rationalised in terms of the different localisation

of the LUMO [54]. Thus, in the case of NNC complexes 35 and 36, the LUMO is

localised on the NNC ligand: the charge transfer occurs in the direction of the

TiO2, to which the NNC ligand is anchored. On the other hand, in the NCN

complex 34, the LUMO is localised on the terpyridine. Electron density thus

moves in the direction of this ligand in the excited MLCT state, remote from the

TiO2, and from where electron injection has low efficiency. Berliguette’s study

considered, in addition, the alternative scenario in which an NCN ligand is used

in conjunction with an ester-substituted terpyridine, where the CT should occur

in the desired direction. However, the long-wavelength absorption tail in such

complexes was found to extend less far into the red—and to be less intense—

than for the corresponding NNC analogues, and so the conclusion is again that

the NNC systems are preferable.



Metal Complexes of Pincer Ligands: Excited States, Photochemistry, and. . .

PF6



(PF6)2

N

N



N



Ru

N



N



PF6



N



N



N



N

Ru



N



PF6

N



Ru



N

N



111



N



N



N



N



N

Ru



N



N

N



CO2H

33



CO2H

CO2H

34



35



HO2C



CO2H

36



Fig. 20 Photocurrent action spectra of 33–36 in a TiO2 solar cell using 0.5 M LiI and 0.05 M I2 in

g-butyrolactone as electrolyte, with the performance of N719 shown for comparison (reprinted

with permission from [52] # 2010 American Chemical Society)



5.2



Iridium(III) Complexes



As seen above, dpyb binds as an NCN-coordinating ligand to Ru(II) and Os(II) ions,

cyclometallating at position 2 of the benzene ring. In contrast, reaction of dpybH

with hydrated iridium(III) chloride gives products in which the ligand metallates at

position 4 of the central ring: the ligand is thus bound in a bidentate fashion, with

one pendant, uncoordinated pyridine ring (N^C binding mode, Fig. 21) [55]. It

seems that metallation at C4 is kinetically favoured over C2. A similar pattern of

reactivity has been observed for palladium(II), at least when Pd(OAc)2 is used as

the metal source [56].



112



G.R. Freeman and J.A.G. Williams

5



5



6



4



6



1



3



1

3



2



N



4



N

M



N



M

N



N



N



N



N



2



N

45



N^C binding mode



N^C^N binding mode

N



N

38



Fig. 21 NCN and NC binding modes of 1,3-di(2-pyridyl)benzene; the structure of 1,3,5-tris

(2-pyridyl)benzene, 38, in which all three C–H bonds of the benzene ring are equivalent; and

the ligand 1,3-bis(N-methyl-benzimidazol-2-yl)benzene, 45, which favours the NCN over the NC

binding mode, even with Ir(III)



Blocking of the C4 and C6 position of dypbH with substituents such as CH3, CF3

or F successfully directs the metallation to position C2, allowing NCN-coordinated

Ir(III) complexes to be obtained. The initial products formed upon reaction with

IrCl3Á3H2O are chloro-bridged dimers of the form [Ir(NCN)Cl(m-Cl)]2 (e.g. 37,

Fig. 22), which can be cleaved by a variety of ligands to generate mononuclear

complexes [57].

Another way to induce the NCN binding mode is to use the C3 symmetric

proligand 1,3,5-tris(2-pyridyl)benzene (38, Fig. 21): all three C–H bonds in this

compound are equivalent, so the question of regiochemistry does not arise.

A representative selection of the NCN-coordinated mononuclear complexes of

varying charge that can be obtained via the intermediacy of the chloro-bridged

dimer, 37, obtained from 1,3-bis(2-pyridyl)-4,6-dimethylbenzene (dpyxH), is

illustrated in Fig. 22. Reaction with terpyridines leads to dicationic complexes

such as [Ir(dpyx)(ttpy)]2+, 39, the Ir(III) analogue of the Ru(II) complex 30

discussed in Sect. 5.1. Dimetallic complexes of the type [{Ir(dpyx)}2(m-tpy-fntpy)]4+ (n ¼ 0–2), 40, comparable to the Ru dimer 31, can be isolated upon reaction

of 37 with back-to-back bridged terpyridines [58]. If a bipyridine is used rather than

a terpyridine, only two of the three chloride ligands around the metal are displaced,

leading to monocationic complexes of the type [Ir(dpyx)(bpy)Cl]+, 41.

Iridium(III) has a higher propensity to undergo cyclometallation than ruthenium

(II): Ir(III) complexes incorporating two or three metallated carbon atoms in the

coordination sphere are common [2], whereas Ru(II) is typically limited to monocyclometallation. In the present instance, further cyclometallating ligands can

readily be introduced. For example, reaction of 37 with 6-phenyl-2,20 -bipyridine

(phbpyH) leads to [Ir(dpyx)(phbpy)]+, 42, a doubly cyclometallated bis-terdentate

complex featuring cis-C2N4 coordination [59]. Such complexes are terdentate

analogues of the well-known class of cationic tris-bidentate iridium complexes of

which [Ir(N^C-ppy)2(N^N-bpy)]+ is the archetypal example [60]. They display

electrochemical properties quite similar to those of [Ru(bpy)3]2+, viz. reversible

first oxidation and reduction waves at accessible potentials (Fig. 23), rendering

them of interest for a variety of applications such as light-emitting electrochemical

cells and photosensitisers for “water splitting” [61].



Metal Complexes of Pincer Ligands: Excited States, Photochemistry, and. . .



113



N

Ir

N



N



[Ir(dpyx)(dppy)]



43



+



N



N

Cl



dppyH2



Ir

N



Ir



AgOTf



N



Cl



N



N



N



no solvent

bpy



ppyH

AgOTf



[Ir(dpyx)(ppy)Cl]



44



no solvent



N



HO



Cl



OH



[Ir(dpyx)(bpy)Cl]+



N



41



Ir

Cl



Cl

Ir



N



N

Cl



+



N



2+



phbpyH



HO



ttpy



37



OH



HO



OH



N



N



Ir

N



N



N



N

Ir



tpy–(F)n–tpy

HO



N



N



OH



+



[Ir(dpyx)(phbpy)]



4+



42



[Ir(dpyx)(ttpy)]2+

N



N

Ir

N



N

N



N

n



N



N



39



Ir

N



N



n=0–2



40



Fig. 22 Structures of a range of Ir(III) complexes containing the NCN-bound dpyx ligand,

obtained through the intermediacy of the chloro-bridged dimer 37



Alternatively, 2,6-diphenylpyridine (dppyH2) can be introduced as a biscyclometallating CNC-binding ligand to generate charge-neutral complexes such

as [Ir(dpyx)(dppy)], 43, terdentate analogues of the much-studied [Ir(ppy)3] class of

compound. Note, however, that whereas [Ir(ppy)3] can exist as either fac or mer

isomers [62], a mer-like configuration of donor atoms is enforced in [Ir(dpyx)

(dppy)]. The two metallated carbon atoms of dppy are forced to adopt positions



114



G.R. Freeman and J.A.G. Williams



Fig. 23 Cyclic voltammogram of 42 in CH3CN at 298 K, at a scan rate of 300 mV sÀ1. The insets

show the redox processes at a series of scan rates (50, 100, 200, 300, 400, and 500 mV sÀ1); the

reversibility of each process is confirmed by the linear dependence of the current on the square root

of the scan rate



trans to one another, and are therefore labilised, due to the high trans influence of

cyclometallated carbon atoms. The result is that [Ir(dpyx)(dppy)] has poor stability

in solution, particularly in the presence of light, under which conditions one of the

two trans-disposed Ir–C bonds is cleaved, and the sixth coordination site probably

occupied by a weakly bound solvent molecule. It is interesting to note that no such

instability is observed for complexes of the type [Ir(dppy)(NNN)]+, despite their

having the same arrangement of trans Ir–C bonds [63]. Evidently, the ligand field

states through which Ir–C dissociation occurs in [Ir(dpyx)(dppy)] are destabilised

by the presence of the third cyclometallated carbon. 2,6-Bis(2-benzimidazolyl)

pyridine has similarly been used recently as a “pseudo-biscyclometallating” analogue of dppy, involving deprotonation of the benzimidazole rings to give a

dianionic ligand [64].

If a bidentate cyclometallating ligand, such as ppy, is used in place of dppy, then

a chloride ligand remains in the sixth site, e.g. [Ir(dpyx)(ppy)Cl], 44. Being trans to

the Ir–C bond, the Ir–Cl does show some lability, but the rate of dissociation is

much slower than that observed for [Ir(dpyx)(dppy)]. Moreover, the chloride can be

replaced by stronger field, acceptor ligands such as CNÀ, which increase the

stability.



Metal Complexes of Pincer Ligands: Excited States, Photochemistry, and. . .



115



Fig. 24 Simplified schematic energy level diagram showing the influence of cyclometallation

on the frontier orbitals in bis-terdentate Ir(III) complexes containing NCN-coordinated

dipyridylbenzene ligands, leading to excited states of different character. A and B represent the

two different ligands [11]



The different classes of complexes that can be generated, as summarised in

Fig. 22, differ greatly from one another in terms of their luminescence properties,

particularly with regard to the quantum yields. The underlying factors for the trends

have been elucidated with the aid of TDDFT calculations, revealing the influence of

the number of cyclometallated carbon atoms on the energies of ligand and metal

orbitals (Fig. 24). Here we present a brief summary of the main points; the reader is

directed to a more comprehensive recent review for further details [65].

The [Ir(NCN)(NNN)]2+ class of complex shows no luminescence at room

temperature, and even at 77 K, emission is barely detectable. [Ir(NNN)2]3+

complexes, in contrast, are modestly luminescent in solution at 298 K (F typically

around 1–3 %) [66, 67]. TDDFT calculations on [Ir(dpyx)(tpy)]2+ reveal that the

lowest-energy excited state has a large degree of NCN ! NNN ligand-to-ligand

charge-transfer (LLCT) character, with the HOMO predominantly localised on the

former and the LUMO on the latter, and with little contribution from the metal

orbitals. Thus, there is a low degree of orbital overlap and, probably, inefficient

SOC, leading to a low radiative rate constant. The situation contrasts with the case

of ruthenium(II), where cyclometallation had a beneficial effect, since the Ru 4d

orbitals are higher in energy, ensuring MLCT character.



116



G.R. Freeman and J.A.G. Williams



The introduction of a second cyclometallating carbon atom into the other ligand,

as in 42, raises the energy of metal-centred orbitals, and reduces the LLCT

character, both of which lead to an augmentation in the radiative rate constant.

Filled metal and ligand orbitals probably become quite comparable in energy in this

case, with the LUMO being localised on the NN part of the NNC ligand, leading to

higher MLCT character to the excited state. Indeed, these complexes emit quite

efficiently in the orange-red region of the spectrum, with quantum yields of around

2–6 % according to the substituents, comparable to [Ru(bpy)3]2+ and [Ir

(ppy)2(bpy)]+ derivatives [59]. Finally, the introduction of a third cyclometallating

carbon atom, as in 43, or an anionic chloride as in 44, raises the energy of the metal

orbitals further, leading to unequivocal MLCT character. Indeed, complex 44 is

very highly emissive in degassed solution at room temperature: Flum ¼ 76%,

t ¼ 1.6 ms, lmax ¼ 508 nm in CH3CN. OLEDs employing 44 or its derivatives as

triplet-harvesting dopants show high efficiencies, competitive with many of the best

iridium emitters reported to date. Haga and co-workers have also obtained similar,

highly efficient emitters in their work using 1,3-bis(N-methyl-benzimidazol-2-yl)

benzene (45, Fig. 21) as an NCN ligand [68]. Interestingly, in that case, no C4/6

blocking groups are required in the central ring: NCN rather than NC coordination

is observed with the parent ligand.

An interesting feature of cyclometallation is that it is accompanied by an increase

in the electron density within the metallated aryl ring, particularly at the position

para to the site of metallation. The propensity of the ring to undergo electrophilic

aromatic substitution reactions is thus enhanced selectively at this position [69].

This feature can be put to use for introducing functionality into a ligand postcomplexation, rather than into the ligand prior to complexation. For example, [Ir

(dpyx)(ttpy)]2+, 39, undergoes bromination under mild conditions (NBS in MeCN at

room temperature) specifically at the para position of the central ring of dpyx.

Naturally, in the case of the bis-cyclometallated complex [Ir(dpyx)(phbpy)]+, 42,

the lateral phenyl ring of the NNC ligand undergoes a comparable reaction. The

modified system [Ir(dpyx)(mtbpy-f-Br)]+ (46, Fig. 25) has been developed as a core

for linear stepwise expansion along the Cdpyx–Ir–Nphbpy axis, through an iterative

sequence of cross-coupling with a boronic acid-substituted substrate (a purely

organic compound or a metal complex), in situ bromination, and a second crosscoupling [59, 70]. The bromination step occurs specifically at the dpyx ligand,

because the corresponding reaction at the phenyl ring of the NNC ligand is blocked

by the presence of the methyl group. Using this strategy, a variety of linearly

elaborated mononuclear complexes and heterotrimetallic systems have been created

in a controlled regiospecific manner. With regard to such multimetallic systems, it

should be noted that bis-terdentate complexes offer structural advantages over those

with bidentate ligands, because the latter, normally having D3 or C2 symmetry, are

chiral, whereas the former (normally D2d symmetry) are not. When two or more

chiral complexes are linked together, mixtures of diastereoisomers are formed

unless the starting mononuclear complexes are resolved first.

It is, however, possible to prepare chiral complexes with terdentate ligands, if the

ligand is desymmetrised, i.e. comprising different lateral groups on either side of the

central ring. An elegant example comes from the group of Haga, who were able to



Metal Complexes of Pincer Ligands: Excited States, Photochemistry, and. . .



117

n+



+

N



N

Ru



(HO)2B

N



N



N



Br



Ru

Pd(0) cat.



N



N



N



46

NBS, MeCN



n+

n+

N



N

Ru



N



N

Ru



Br



N



N



N



N

B(OH)2



Pd(0) cat.



Fig. 25 Sequential cross-coupling—bromination—cross-coupling applied to complex 46 for the

elaboration of bis-terdentate iridium(III) complexes, relying on the activation of the 40 -position of

the dpyx ligand to electrophilic substitution reactions. The shaped spheres represent either chargeneutral organic units (hence n ¼ 1) or metal complexes, in which case, n depends on the charge of

the complexes introduced



resolve complex 47 into its two enantiomers, as confirmed by their equal but opposite

circular dichroism spectra (Fig. 26, top) [71]. More unusually for transition metal

complexes, circularly polarised luminescence spectra could also be recorded (Fig. 26,

bottom).



5.3



Platinum(II) Complexes



5.3.1



Pt(dpyb)Cl and Derivatives



Reaction of dpybH with K2PtCl4 gives the NCN-coordinated complex Pt(dpyb)Cl,

48 [56]. In contrast to the reaction with Pd(OAc)2 or IrCl3, there is no evidence of

competitive metallation at C4/6 in the case of platinum(II).



N

N



Pt



N



N



Pt



N



Pt



Cl



Cl



Cl



48



49



50



N



118



G.R. Freeman and J.A.G. Williams

40



Δε / M-1cm-1



20



0



-20



-40

250



300



350



400



500



450



550



Wavelength / nm

3



CD / mdeg



2

1

0

-1

-2

-3

450



500



550



600



650



Wavelength / nm

Fig. 26 Circular dichroism spectra (top) and circularly polarised luminescence spectra (bottom)

of the two enantiomers of complex 47 resolved by HPLC on a chiral column. Solvent ¼ CH2Cl2

(reprinted with permission from [71] # 2009 Royal Society of Chemistry)



Pt(dpyb)Cl displays intense green luminescence in solution at room temperature,

Flum ¼ 0.60 and t ¼ 7.2 ms in deoxygenated dichloromethane [72]. The superiority compared to the equivalent non-metallated complex [Pt(tpy)Cl]+, which is

essentially non-emissive under these conditions, is extraordinary. This can be

attributed, at least in part, to the stronger ligand field of the metallated system

ensuring that deactivating d–d excited states are shifted to higher energy, similar to

the beneficial effect discussed earlier for ruthenium(II) [73]. However, the quantum

yield and lifetime of this complex are also an order of magnitude superior to those

of the isomeric complex Pt(phbpy)Cl, 49, containing an NNC rather than a NCN

ligand [74], and of the related bidentate complex Pt(ppy)(ppyH)Cl, 50, which has

the same local coordination sphere comprised of two pyridine rings: one

cyclometallated carbon atom and one chloride ligand [75]. The greater rigidity

associated with binding of a metal to a terdentate ligand compared to a bidentate

one could explain the latter point: the instability of square planar complexes with



Metal Complexes of Pincer Ligands: Excited States, Photochemistry, and. . .



119



Fig. 27 Showing how the emission spectra of the Pt(dpyb)Cl class of complex can be tuned from

that of the parent 48 (third from left), either to longer wavelengths through the introduction of

electron-donating substituents at the 4-position, or to shorter wavelengths using electronwithdrawing groups at this position/substituents in the pyridyl rings. Spectra are recorded in

CH2Cl2 at 298 K; lex ¼ 400 nm; intensities have been normalised



bidentate ligands with respect to distortion from D4h to D2d symmetry in the excited

state has long been recognised as a potential sink for the excited-state energy [3, 12].

As far as the pair of terdentately bound complexes are concerned, it has been noted

˚ vs. 2.05 A

˚ ) in the NCN

that the Pt–C bond is significantly shorter (ca. 1.90 A

complex than in the NNC analogue, which may ensure a stronger ligand field and

hence displacement of the deactivating d–d states to higher energy [72]. However, a

recent comparison of these systems by TDDFT suggests that the key point is the

greater rigidity of the former, with a lower degree of reorganisation in the excited

state compared to the ground state [76].

The emission energy of complexes of this type can be tuned over a wide range,

without significantly compromising the quantum yields, by introducing substituents

into the cyclometallated aryl ring or into the pyridine rings. Electron-releasing

substituents in the central 4-position of the benzene ring lead to increasingly redshifted emission according to their electron-donating ability, e.g. lmax increases in

the order R ¼ CO2Me < H < mesityl < Me < 2-pyridyl < 4-tolyl < 4-biphenylyl

< 3,4-dimethoxyphenyl < 2-thienyl (Fig. 27) [77]. There is a good correlation

between the emission energy and the oxidation potentials Epox of the complexes,

with a slope of 4,900 cmÀ1VÀ1, whereas the reduction potentials vary little with the

4-substituent. The obvious inference is that the substituent influences primarily the

HOMO, such that the observed red shift with increasing electron-donating ability

reflects an increase in the HOMO level and a LUMO that remains essentially

unchanged. TDDFT calculations support this picture, revealing a major contribution to the HOMO but not to the LUMO (Fig. 28a) [78]. The LUMO, in contrast, is

localised predominantly on the pyridyl rings, and it may be noted that the 40 position makes no contribution to the HOMO. Thus electron-donating substituents

in this position are expected to increase the LUMO level without significantly



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

1 Ruthenium(II) and Osmium(II) Complexes

Tải bản đầy đủ ngay(0 tr)

×